On Freitag, 19. Oktober 2007, Rob Church wrote:
On 19/10/2007, Ralph Hulslander rph@rovenet.com wrote:
Well I do not get it.
While the functioning definitely is interesting (I have very little wiki knowledge so interesting means a lot to me).
I do not understand the relevancy of the questions. Is this a "quiz" why should I spend my time answering questions about a subject of no interest to me?
As I understand it, a broad application might be importing a Wikipedia dump into the site and using user feedback to help automated populating the semantic data store. The questions are thus the extraction tool asking for some human verification as to whether it's ascertained the correct relationship between two entities, e.g. "is X a place in Y?". This is far more effective, and less boring, than asking new visitors to populate a structured set, which while not too hard, is a higher learning curve than editing a free-form text page.
Obviously, this would be of no interest to people helping to develop the semantic interpretation of data, or anyone not interested in the mission of that particular project, e.g. Wikipedia, where the mission is to create a free encyclopaedia.
Yes, that were exactly our ideas :-) Of course one could also use the extension in other ways, e.g. to verify some other robot input that is not so certain, or even to do something completely different. We first need to make it more stable and then we will leave this to the creativity of all of you ...
Ralph also suggested "questions relative to the content of the page". Indeed we have foreseen ways of displaying questions only on pages that they relate to. But since our test wiki has 1.5million pages but only a couple of thousands of questions, this would make it rather unlikely for you to ever find a question ;-) If questions on pages are altogether undesired, and one could even show questions only on the special page, so that just interested users ever see them.
Markus